What this can answer
- How many fatal crashes involved each mode in a state in the selected year.
- Which cells met the small-cell publication floor.
- Which official release and transformation produced the counts.
Explore · compare · cite
One evidence set, five ways to inspect it
The national map and every analytic view are driven by the same 306 reviewed state-by-mode cells. A selection in one view follows you into the others.
State loads the five-year profile. Year and visualization focus drive the national views; publication status filters the matrix and ledger. The complete ledger has its own mode filter below.
Loading the reviewed artifact…
Whole-country view
Fatal-crash burden across the United States
Select a road-user mode to read the same reviewed counts geographically. Color is scaled within that mode; hatching means suppressed or zero.
Keyboard: Tab enters the map once. Arrow keys, Home, and End move through jurisdictions in alphabetical order; Enter or Space selects.
Lower published count
Higher published count
Suppressed or zero
Loading state evidence…
Publication landscape
State × involved-mode matrix
Each column has its own scale. Hatched cells are not published and are never treated as zero.
Keyboard: after the matrix scroll region and mode filters, Tab enters the cell grid once. Arrow keys move between cells; Home and End move to the first and last mode; Enter or Space selects.
| State | Involved road-user mode | |||||
|---|---|---|---|---|---|---|
| Loading the reviewed artifact… | ||||||
Published burden
Rank one involved mode
Only numeric cells are ranked. This is burden, not exposure-normalized risk or safety.
Keyboard: Tab enters the ranking once. Arrow keys, Home, and End move through states; Enter or Space selects.
Count relationship
Compare two involved modes
States appear only when both cells are published. Missing values stay outside the plot.
Keyboard: Tab enters the plot once. Arrow keys, Home, and End move through states; Enter or Space selects.
Open the accessible comparison table
| State | Involved road-user mode | |
|---|---|---|
| Loading the reviewed artifact… | ||
Side by side
Compare two states without adding modes
Every mode keeps an independent scale because one crash may contribute to multiple rows.
Compare years with care. FARS person-type coding uses one reviewed semantic regime for 2020–2021 and another for 2022–2024. Each result stays pinned to its annual mapping contract, so a cross-regime change can reflect coding as well as outcomes.
Five-year view
Selected-state profile, 2020–2024
Each year is kept separate. The coding seam marks two reviewed FARS person-type regimes, so values on opposite sides should not be read as one continuous series.
Choose a state above to load its five-year profile.
| Year | Motor-vehicle occupant | Motorcyclist | Pedalcyclist | Pedestrian | Other road user | Unknown mode |
|---|
Carry the caveat with the number
Evidence brief
Save up to four states. The printable brief keeps the year, mode, publication status, source, and “counts are not risk” limit attached.
No evidence saved yet. Select a state, then add it from the inspector.
Counts are not risk. They do not account for population, trips, miles, or exposure. Modes can overlap in one crash, so mode counts are not additive and do not imply fault or causation.
Auditable rows
Complete state × mode evidence ledger
A published count is the number of distinct fatal crashes with at least one person in that mode. A non-published cell is labeled “suppressed or zero”—never zero.
| Year | State | Involved road-user mode | Publication status | Fatal-crash count |
|---|---|---|---|---|
| Loading the reviewed artifact… | ||||
Trace the evidence
Source and deployment provenance
The public JSON and release index carry source, mapping, semantic-regime, small-cell publication, and accounting proof. Deployment files bind the published page back to an exact source commit.
What this cannot answer
- Risk, rate, fault, causation, or a safety ranking.
- Near misses, nonfatal crashes, counties, streets, or intersections.
- Whether a non-published cell is zero or suppressed.
Five years, with publication status made explicit
Official source archives are pinned for 2020–2024. Only years whose reviewed public artifacts appear in the release index can be selected; unavailable years are never silently treated as zero.
- 2020Source pinnedState counts published
- 2021Source pinnedState counts published
- 2022Source pinnedState counts published
- 2023Source pinnedState counts published
- 2024Source pinnedState counts published
- Official source
- NHTSA FARS
- Source revision
—- Semantic regime
—- Mapping versions
—- State-code system
—- Annual contract
—- Map boundaries
- U.S. Census Bureau 2024 Cartographic Boundary Files
- Boundary file SHA-256
—